Jaw movements in the production of vowel sequences were analyzed by means of the framework of dynamic model. An instrument to observe the jaw movements using an optical spot position sensitive detector was developed. Utilizing this instrument, the jaw movements in the production of meaningless three-mora Japanese vowel sequence words were observed. As the basic analysis, the jaw opening for each vowel and the time constant for each transition at the extremely slow speaking rate were analyzed. Afterwards, the target value of jaw opening at the fast speaking rate was estimated. In this study, the time function of the observed jaw movement was approximated by the response of the critically damped linear second order system driven by the step function, whose amplitude represents the target value of the jaw opening for each vowel. Results obtained in this study suggest that the target value of jaw opening for each vowelshould be different depending on the speaking rate and the word-internal vowel position. However, for the tendency concerning the difference in the target values caused by the adjacent vowels, common characteristics for two subjects of this study were not observed.

View full abstract

In this paper, the complexity of wave motions has been formulated in order to more detailedly than previously consider the complexities of reverberant sound fields, and the difference of complexity between two separate points has been described in an attempt to find a measure of diffuseness in these fields. The reverberant fields are assumed to be sound fields where random plane waves propagate two- or three-dimensionally. Entropy H (=h₂-h₁)and energy-entropy product K are used to define the complexity of wave motions, where h₁, h₂ are entropies by sound pressures p₁, p₂, respectively. Then H is compared with the cross-correlation coefficient R calculated from same sound pressures. Furthermore, K is obtained by considering energy dissipation with propagation, and its properties are discussed as well as those of H. The results show that the entropy of sinusoidal wave is in a complementary relationship with the cross-correlation coefficient; H=1-R, and that it may be convenient to measure since the cross-correlation coefficient is inseparable into two entropies. Either the entropy or the energy-entropy product is also found to be required at audio frequencies as a measure of the complexity of wave motions (i.e. approximately K=A₀H, A₀=const. for this case). Because of the lack of a quantity giving degrees of directional distribution, our introduced measure is as yet insufficient to accurately measure the diffuseness of sound fields, but it may be fairly near to the corrected measure.

View full abstract

This paper describes the speaker independent isolated word speech recognition method developed for telephone speech response systems. To recognize speech, input utterances are first frequency analyzed by 19 channel BPFs. The frame cycle used is 8ms. Then the analyzed data undergo logarithmic conversion, normalization of voice chords sound source characteristics by least squares approximation line and time normalization by linear companding to 32 frames. The speech patterns thus obtained undergo pattern matching with multiple reference patterns generated separately for male and female speakers in advance. In applying this recognition method, it is necessary to optimize the reference patterns so that the speech can be correctly recognized in spite of the difference of formant frequencies, the differences in individual speaker's habits, the variations of phonetic positions, non-vocalization, and slight segmentation errors. To evaluate the performance of this recognition method, voices of about 2, 000 persons were recorded through long distance telephone lines. A 16 Japanese words vocabulary was used. A total of 256 male and female reference patterns were generated using the training voice data of about 570 persons. The speech recognition accuracy of this method in recognizing non-training voice data was 97.8 %.

View full abstract

Local chemical dissolution, i.e. chemical lithotresis, of calcium salt crystal (gypsum) in EDTA solution was made by irradiating the ultrasonic wave locally on the surface of gypsum. The dissolution, however, was hindered by the macrobubbles (0.1 mm order in diameter) on the surface of gypsum. It seems that the macrobubbles contained in the solution and crystal have grown up into the macrobubbles on the surface of gypsum byirradiating the ultrasonic wave. These macrobubbles gathered to the surface of solution when the ultrasonic irradiation stopped, so that the intermittent ultrasonic wave was used in this study. It takes about 0.5 s in order to remove the macrobubbles almost completely from the surface of gypsum. The rate of the chemical lithotresis is proportional to the power of ultrasonic wave under this experimental condition unless the macrobubbles exist on the surface of gypsum.

View full abstract

This paper reports and discusses the results of a round robin test in Japan for the loudness of impact sound, which aims to establish how to evaluate the loudness, the noisiness and the annoyance of impact sounds and to describe them. As impact sounds which are encountered in daily life are characterized by high peak levels and short durations, the stimuli with such characteristics were prepared for the test: i.e. they had a short rising part (1 ms for 20 dB change), a steady part (0-100 ms) and a decaying part (30-300 ms). In this study the single burst of impact sound were presented diotically through headphones, and the method of constant stimuli was used. It is shown as the resultof the test that the sound energy level or the maximum output of an r.m.s. circuit with a time constant of 125 ms 1 s is suitable for the evaluation of stimuli. When eitherof these quantities is used, however, impact sounds are slightly louder (equivalent to 1-2 dB) than the burst of 200 ms long.

View full abstract